ObjectiveTo compare the performance of different automatic evaluation methods for detecting pattern hallucinations generated by large language models （LLMs） in traditional Chinese medicine （TCM） syndrome differentiation and to identify a strategy with higher overall discriminative performance against expert manual judgment. MethodsA standardized TCM syndrome-differentiation dataset containing 598 cases was constructed from case reports published in Chinese core journals indexed by Peking University Core Journals or Chinese Science Citation Database. Each of the 598 cases was submitted to five Chinese LLMs， with each model generating one syndrome-pattern output per case， yielding 2990 outputs in total. The 598 outputs generated by Qwen3-Next-80B-A3B-Thinking were independently annotated for syndrome-pattern hallucinations by two licensed TCM physicians with intermediate or higher professional titles. Disagreements were adjudicated by a third licensed TCM physician with a senior associate professional title and more than 10 years of clinical experience. The resulting consensus annotations served as the reference labels for training and validating the automated evaluation method. Based on these reference labels， three categories of automated evaluation methods were developed and compared following a progressive strategy from text-level semantic similarity， to syndrome-element structural consistency， and finally multi-feature fusion. Method A employed sentence-embedding models based on semantic similarity between pattern texts. Method B adopted a rule-based threshold method using weighted Jaccard coefficients of disease-location and disease-nature syndrome elements. Method C utilized a machine-learning classifier integrating syndrome-element matching scores， missing and redundant syndrome-element counts， and multiple semantic-similarity features. Five-fold cross-validation was used to evaluate the discriminative performance of each method against manual judgment. The optimal method was then applied to uniformly assess the pattern hallucination rates of five LLMs. ResultsIn method A， text2vec-large-chinese showed best overall performance， with an area under the receiver operating characteristic curve （AUROC）， area under the precision-recall curve （PR-AUC）， F1-score， sensitivity， and specificity of 0.849， 0.884， 0.794， 0.801， and 0.691， respectively. In method B， the corresponding metrics of the rule-based syndrome-element Jaccard threshold method were 0.848， 0.851， 0.817， 0.845， and 0.684， respectively； those of the LLM-based syndrome-element Jaccard method were 0.872， 0.906， 0.795， 0.764， and 0.780， respectively. In method C， the multilayer perceptron （MLP） classifier showed the best overall performance， with AUROC， PR-AUC， F1-score， sensitivity， and specificity of 0.933， 0.951， 0.871， 0.857， and 0.846， respectively. Under the unified evaluation framework combining DeepSeek-R1 syndrome-element extraction and MLP classifier， the estimated pattern hallucination rates of the five LLMs ranged from 56.0% to 77.4%. ConclusionAmong the three automatic evaluation methods， Method C integrating syndrome-element discrepancy features and text semantic features using an MLP classifier showed the highest overall discriminative performance against manual judgment， and is more suitable for evaluating syndrome hallucinations in LLM-based TCM syndrome differentiation tasks.

Objective To evaluate the clinical efficacy of haloperidol combined with auricular point pressing therapy in children with tic disorders.Methods A total of 120 pediatric patients diagnosed with tic disorders at the Department of Pediatric Rehabilitation,Cangzhou Hospital of Integrated Traditional Chinese and Western Medicine,Hebei Province from March 2022 to March 2024 were enrolled.Participants were randomly divided into an observation group(n=60)and a control group(n=60)using a random number table.The control group received haloperidol alone,while the observation group received additional auricular point pressing therapy.Treatment duration was 12 weeks.After 3 months,clinical efficacy was assessed by comparing:Yale Global Tic Severity Scale(YGTSS)scores,serum 5-hydroxytryptamine(5-HT)and dopamine(DA)levels,immune function markers(CD3+,CD4+),and incidence of adverse reactions between groups.Results(1)The overall efficacy rate in the observation group was 91.67％(55/60),while that in the control group was 76.67％(46/60).The efficacy of the observation group was superior to that of the control group,with a statistically significant difference(P＜0.05).(2)After treatment,the YGTSS scores of children in both groups was significantly improved(P＜0.05),and the observation group showed a significantly greater improvement in YGTSS scores than the control group,with a statistically significant difference(P＜0.05).(3)After treatment,the levels of 5-HT and DA in both groups of children were significantly improved(P＜0.05),and the observation group showed a significantly greater improvement in 5-HT and DA levels than the control group,with a statistically significant difference(P＜0.05).(4)After treatment,CD3+and CD4+levels were significantly improved in both groups(P＜0.05),and the observation group showed a significantly greater improvement in CD3+and CD4+levels than the control group,with a statistically significant difference(P＜0.05).(5)There was no statistically significant difference in the incidence of adverse reactions between the observation group and the control group,with a statistically significant difference(P＞0.05).Conclusion The combination of haloperidol and auricular point pressing therapy significantly improves clinical symptoms and immune function in children with tic disorders,demonstrating both efficacy and safety.

Objective: To compare the difference in somatic gene mutation of PTC subtypes between 114 patients with papillary thyroid carcinoma (PTC) and The Cancer Genome Atlas (TCGA) database. Methods: Totally 114 PTC patients admitted to The First Affiliated Hospital of Nanjing Medical University were recruited. The 18 hotspot genes associated with thyroid cancer were detected in thyroidectomy specimens were using next generation sequencing. PTC data were downloaded from the TCGA database in the cBioPortal website, and the difference in the somatic gene mutation was compared between 114 PTC patients and the TCGA database Results: The 114 PTC patients included 73 women (64.04%) and had a mean age of (39.23±13.18) years. The prevalence of BRAF V600E (66.67% vs. 48.68%), TERTp (3.51% vs. 0.41%), PDGFRA (1.75% vs. 0%), PTEN (3.51% vs. 0.41%) and TP53 gene mutations (4.39% vs. 0.61%) was significantly higher among the 114 PCT patients than in the TCGA database (P<0.05). The prevalence of BRAF V600E (80.88% vs. 54.99%), TP53 (7.35% vs. 0.57%) and TSHR gene mutations (2.94% vs. 0%) was significantly higher in classical PTC(CPTC) patients than in the TCGA database, and the prevalence of BRAF V600E (36.84% vs.13.86%) and TERTp gene mutations (10.53% vs. 0%) was significantly higher in follicular variant PTC (FVPTC) patients than in the TCGA database. According to the American Thyroid Association Risk Stratification of Thyroid Cancer Recurrence, the prevalence of BRAF V600E and TP53 gene mutations was 77.14% and 8.57% among moderate-risk CPTC patients, the prevalence of BRAF V600E gene mutation was 27.27% among low-risk FVPTC patients, and the prevalence of TERTp gene mutation was 33.33% among moderate-risk FVPTC patients, which were all higher than in the TCGA database (55.10%, 0%, 3.28%, and 0%, respectively; P<0.05). Conclusion There are significant differences in the type and rate of somatic gene mutations between 114 PTC patients and the TCGA database.

Objective:To investigate interleukin (IL)-36 expression in patients with myasthenia gravis (MG), and to study the modulatory function of IL-36 on regulatory T cells (Tregs) and Th17 cells in MG patients.Methods:Fifty-one MG patients (MG group) and 25 healthy controls (control group) were enrolled in this study in Xinxiang Central Hospital between July 2016 and August 2021. Peripheral blood was collected. Plasma and peripheral blood mononuclear cells (PBMCs) were isolated. Plasma IL-36α, IL-36β, IL-36γ, IL-36RA, IL-35, and IL-17 levels were measured by enzyme-linked immunosorbent assay. The percentages of Tregs and Th17 cells were measured by flow cytometry. Forkhead box protein P3 (FoxP3) and retinoid-related orphan receptor gamma t (RORγt) mRNA expressions were measured by real-time polymerase chain reaction. PBMCs or purified Tregs from MG patients were stimulated with recombinant IL-36β (5 ng/ml). Changes of Tregs and Th17 cell percentages, IL-35 and IL-17 secretions, FoxP3 and RORγt mRNA expressions, as well as immunosuppressive activity of Tregs were analyzed.Results:There were no statistically significant differences of IL-36α, IL-36γ, or IL-36RA between the control group and the MG group (all P>0.05). IL-36β level was notably higher in the MG group compared with the control group [(73.43±13.91) pg/ml vs (60.91±12.65) pg/ml, t=3.79, P<0.001]. Treg percentage [(4.67±1.33)% vs (6.32±1.81)%, t=4.48, P<0.001], IL-35 [(50.06±7.93) pg/ml vs (65.37±8.90) pg/ml, t=7.59, P<0.001] and FoxP3 mRNA expression (1.03±0.14 vs 1.57±0.46, t=7.78, P<0.001) was lower, while Th17 cell percentage [(1.05±0.15)% vs (0.94±0.21)%, t=2.61, P=0.011], IL-17 [(40.61±13.13) pg/ml vs (33.09±11.48) pg/ml, t=2.44, P=0.017] and RORγt mRNA expression (1.26±0.16 vs 1.03±0.13, t=6.08, P<0.001) was higher in the MG group ( P<0.05). There were no statistically significant differences of above indices between different genders, onset ages, afflicting with thymoma, or different Osserman types (all P>0.05). There were no statistically significant correlations between above indices and quantitative myasthenia gravis (QMG) score (all P>0.05). Recombinant IL-36β stimulation did not affect PBMCs proliferation in MG patients ( P=0.248), and reduced Tregs percentage [(3.05±0.66)% vs (4.18±1.07)%, t=4.23, P<0.001], IL-35 secretion [(48.12±10.93) pg/ml vs (56.96±13.73) pg/ml, t=2.36, P=0.023] and FoxP3 mRNA expression (0.99±0.17 vs 1.18±0.13, t=4.01, P<0.001), but did not affect Th17 cell percentage, IL-17 secretion or RORγt mRNA expression (all P>0.05). Recombinant IL-36β stimulation inhibited immunosuppressive activity of Tregs, which presented as enhanced cellular proliferation [(0.83±0.12)×10 5vs (0.69±0.15)×10 5, t=3.02, P=0.005] and reduced IL-35 secretion [(28.71±10.08) pg/ml vs (37.12±10.47) pg/ml, t=2.39, P=0.023]. Conclusion:Increased IL-36β contributed to the regulation of Tregs/Th17 cell balance probably through inhibition of Tregs function in MG patients.

Objective:To investigate the characteristics of regional brain functional centrality(DC)in patients with levodopa-induced dyskinesia(LID)and to explore the pathogenesis of LID.Methods:A total of 33 PD patients with LID(PD-LID), 41 PD patients without LID(PD-nLID)and 37 healthy controls from the First Affiliated Hospital of Nanjing Medical University were enrolled in this study.Differences in DC among the three groups were compared and the correlation between Z-DC values of the brain regions with differences and the scores of the involuntary movement scale(items 1-7)was analyzed.Results:Compared with Controls, PD-LID patients showed increased DC in the right amygdala(extending to the right globus pallidus)(MNI: x=30, y=-3, z=-18, t=4.00, P<0.05 after AlphaSim correction)and in the right postcentral gyrus(MNI: x=57, y=-9, z=39, t=-3.59; MNI: x=42, y=-33, z=57, t=-4.23, P<0.05 after AlphaSim correction)and reduced DC in the right superior parietal lobule(MNI: x=24, y=-51, z=72, t=-3.95, P<0.05 after AlphaSim correction).Compared with the PD-nLID group, the PD-LID group showed increased DC in the right globus pallidus(MNI: x=30, y=-12, z=-3, t=3.09, P<0.05 after AlphaSim correction).DC changes in the right globus pallidus were positively correlated with AIMS score( r=0.482, P=0.004). Conclusions:The enhancement of DC function in the right globus pallidus may be closely related to the onset and severity of LID.

Objective:To observe the characteristics of resting-state brain activity in Parkinson disease (PD) patients with peak-dose dyskinesia, and to explore its pathogenesis.Methods:From March 2017 to November 2019, totally 27 PD patients with peak-dose dyskinesia (dyskinetic group), and 29 PD patients without dyskinesia (non-dyskinetic group) treated in the First Affiliated Hospital of Nanjing Medical University and 27 healthy controls from the community were recruited.Resting-state functional magnetic resonance imaging (rs-fMRI) and clinical scale data were collected.SPSS 26.0 software and REST software were used for data analysis.The whole brain amplitude of low-frequency fluctuation (ALFF) values were compared among the three groups.Correlation analysis was performed between ALFF values of the significant brain regions and the scale scores.Finally, receiver operating characteristic (ROC) curve was used to evaluate the efficacy of ALFF values of significant brain regions in identifying three groups of subjects.Results:The peak-dose dyskinetic group showed decreased ALFF in right inferior frontal gyrus(MNI: x=36, y=21, z=30; x=36, y=18, z=30)and increased ALFF in right supplementary motor area (MNI: x=9, y=0, z=69; x=6, y=-3, z=72)(all P<0.05, Alphasim correction) compared with non-dyskinetic group and healthy controls.ALFF value in right inferior frontal gyrus was negatively correlated with unified dyskinesia rating scale (UDysRS) scores ( r=-0.47, P=0.018). The ALFF value of the right inferior frontal gyrus was more effective in identifying peak-dose dyskinetic patients from non-dyskinetic patients and healthy controls, and the area under the curve of right inferior frontal gyrus were 0.881 and 0.787 (both P<0.01), respectively. Conclusion:Abnormal spontaneous brain activity in right inferior frontal gyrus and right supplementary motor area can be the neurobiological basis of peak-dose dyskinesia in PD patients.The severity of peak-dose dyskinesia is associated with abnormal brain activity of right inferior frontal gyrus.The ALFF value of right inferior frontal gyrus is a potential imaging marker for identifying peak-dose dyskinetic patient.

Objective To investigate the diagnostic value of pyruvate kinase M2 ( PKM2) gene expression in papillary thyroid carcinoma ( PTC). Methods Quantitative real-time PCR ( RT-qPCR) was used to detect the expression of PKM2 mRNA in benign thyroid nodules, PTC, and normal thyroid cells around nodules of fine-needle aspiration (FNA) specimens. Immunohistochemistry ( IHC) was used to detect the expression of PKM2 protein in thyroid tissue after thyroidectomy. The receiver operating characteristic curve was constructed to evaluate the diagnostic value of PKM2 in PTC. Results The expression of PKM2 mRNA was detectable in FNA specimens of thyroid nodules,higher in PTC than those in normal thyroid tissue and benign thyroid nodules (P<0.01). PKM2 expression level was correlated with diameter of PTC ( P<0.05) , but had no correlation with lymph node metastasis, BRAFV600E mutation, and American Joint Committe on Cancer( AJCC) stage ( P>0.05) . The expression level of PKM2 mRNA in FNA specimens of thyroid nodules was paralleled with the expression level of PKM2 protein in postoperation pathological tissues. The accuracy, sensitivity, specificity of PKM2 gene in the diagnosis of PTC were 62.8％, 46.9％, and 95.7％, respectively. The accuracy and sensitivity of PKM2 combined with BRAFV600E were increased to 87.6％and 83.7％. Conclusion Detection of PKM2 gene in FNA specimens is highly specific in the diagnosis of PTC, making it a valuable molecular marker for preoperative diagnosis. The combination of PKM2 and BRAFV600E detection shows a higher diagnosis efficiency.

Objective: To explore the association between the efficacy of peg-IFN and the complexity of TP and RT regions of hepatitis B virus (HBV) in chronic hepatitis B. Methods: Patients with HBeAg positive, HBV DNA positive chronic hepatitis B were given peg-interferon 180 μg once a week for subcutaneous injection, and baseline information was collected from baseline and after 12 weeks’ treatment. The baseline HBV DNA TP and RT fragments were amplified, database, high-throughput sequencing, and the average genetic distance calculation. Results: Data of 108 patients were analyzed by logistic regression. RT area fragment Markov distance and TP area fragment Shannon quotient for HBV DNA response were calculated. ALT level is good for HBeAg response. HBsAg level is bad for HBsAg response. Conclusions The complexity of the baseline TP and RT regions may be associated with the efficacy of peg-interferon therapy for CHB.

Objective: To investigate the differences in frequency and function of natural killer cells (NK) between chronic hepatitis B (CHB) and acute hepatitis B (AHB). Methods: Patients with AHB and those with CHB in immune active (IA) phase were enrolled. The frequencies of NK, CD56^dimNK, CD56^brightNK and the expression of functional molecules IFNAR2 and NKp46 on the surface of NK cells were detected respectively among patients with CHB in IA phase, patients with AHB, and those recovered from AHB. At the same time, their correlations with ALT, HBV DNA and HBV markers were analyzed. Results: Between IA and AHB, the frequencies of NK cells and NKp46^dim NK cells in AHB cases were significantly lower than those in IA cases, but the frequency of NKp46^high NK cells in AHB was higher than that in IA. For patients who recovered from AHB, the frequency of NK cells and NKp46^dim NK cells increased; the varied ranges of frequencies of CD56^dimNK, IFNAR2⁺ NK and NKp46⁺ NK cells were on the rise, while the frequency of NKp46^high NK cells decreased after the recovery from AHB, and the varied ranges of CD56^brightNK and IFNAR2MFI, NKp46MFI decreased. In AHB, HBVDNA loads were positively correlated with ALT levels. Before and after the recovery of AHB: ΔHBV DNA and ΔALT, Δ NK/LY (%) were positively correlated; ΔALT and ΔNKp46^highNK/NK(%), ΔNKp46MFI, ΔIFNAR2MFI were positively correlated. Conclusions In CHB immune active phase, the activity of peripheral blood NK cells was too weak to remove the virus, but NK cells play an important role in eliminating the viruses and mediating liver tissue inflammation in AHB.

Objective: To investigate the change of hepatitis B surface antibody (HBsAb) titer and its long-term protection and infection rates between 1 and 3-year-old children whose mothers were chronic hepatitis B pregnant woman with HBeAg positive and high viral load after successful blocking of mother-to-child transmission. Methods: One-year-old children whose mothers were hepatitis B surface antigen (HBsAg) and hepatitis B e antigen (HBeAg) positive, with HBV DNA≥10⁶IU/ml were enrolled, then were followed up till 3 years old, and tested the five serological markers of hepatitis B and biochemical parameters at the age of one and three years respectively, and analyzed HBsAb titer, positive rate, negative rate and infection rate of 1 to 3-year-old children without enhanced vaccination; meanwhile, data of HBsAb titers at the age of 7 months were collected HBsAb titer, positive rate, and negative rate were analyzed. Results: Totally 264 1-year-old children were enrolled into the study, including 178 children without enhanced vaccination between seven months and 1 year of age, and 114 children without enhanced vaccination between 1 year and 3 years of age. Our result showed that there were no infected children at the age between 1 and 3 years. HBsAb titer decreased from 7 months to 1 year old and dropped from 1 000 IU/L to 509.43 IU/L (P<0.05), and the antibody was still protective. From 1 year to 3 years old, HBsAb titer dropped from 466.72 IU/L to 67.3 IU/L (P< 0.05); at the age of 3 years, 60.52 % children were either weakly positive or negative, but still protective, but significantly less than those who had the reinforced vaccination. As a result , the children without the enhanced vaccination between 1 and 3 years of age were still at high risk. Conclusions If the antibody was protective at 7 months, children were not easily infected between 1 year and 3 years of age. At the age of 3, the antibody dropped to low or no responsive levels, and the children were still at high risk. It is necessary to take protective measures and supplement the vaccine.